How accurate is echocardiographic measurement of patent ductus arteriosus?

Summary

Objective:
To compare measurements of internal size of patent ductus arteriosus on transthoracic echocardiography with angiography, serving as the gold standard, in infants and children.

Methods:
Retrospective analysis of data from paediatric patients undergoing transcatheter patent ductus arteriosus device closure in a tertiary paediatric cardiac centre between March 2013 and March 2014 was performed. Internal size measurements of patent ductus arteriosus were performed retrospectively on recorded echocardiographic and angiographic images independently on 2 separate occasions by two observers on a mutually agreed image.

Results:
In total 47 patients with available adequate echocardiographic and angiographic images were identified. Median age of patients was 7.5 years (range 8 months – 15 years) and average weight was 12.5 kg (SD 7.5 kg). Intra-observer correlation coefficient was 0.5 (95% confidence interval 0.25–0.69) and inter-observer correlation coefficient was 0.45 (95% confidence interval 0.18–0.65) for echocardiographic measurements and 0.75 (95% confidence interval 0.61–0.85) and 0.83 (95% confidence interval 0.71–0.9) for angiographic measurements, respectively. The Pearson’s correlation coefficient was 0.275 and interclass correlation coefficient was 0.247 between echocardiographic and angiographic measurements of ductal internal size, by a single observer, suggesting poor correlation and agreement. The measurements of the same duct by the same observer from echocardiogram and angiogram were significantly different (p<0.0001).

Conclusion:
Based on our data, angiographic measurement considered the gold standard for assessment of patent arterial ductal size does not correlate well with echocardiographic measurements. This may have implications for treatment strategies based on echocardiographic ductal size measurements only.

KEY WORDS:
patent ductus arteriosus, echocardiography, angiography, infants, children

INTRODUCTION

Patent ductus arteriosus (PDA) is a postnatal remnant of fetal communication between systemic and pulmonary circulation. The incidence PDA has been reported to be around 1 in 2000 live-births in those born at term [1, 2]. This accounts for 5–10% of all congenital heart disease. However, approximately 65% of infants born at less than 26 weeks of gestation will have PDA detectable early postnatally [3]. Left to right shunt across PDA has been associated with increased risk of bronchopulmonary dysplasia, necrotizing enterocolitis, intraventricular haemorrhage and death in this population [4]. Medical or surgical closure of large PDAs in extremely premature infants has been routine practice in many neonatal centres. In most so far published data, internal diameter of PDA measured on transthoracic echocardiography (TTE), has been described as an important factor in determining the need for therapeutic intervention on PDA in premature infants [5, 6]. To our knowledge, internal diameter echocardiographic measurement of PDA has not been validated against another imaging technique in paediatric patients in this century. In our study, we compared internal diameter of PDA measured on TTE with that measured on angiogram performed at cardiac catheterisation (considered gold standard for PDA size) in an attempt to validate accuracy of PDA size measurement on TTE.

METHODS

Retrospective analysis of hospital records in paediatric patients undergoing transcatheter PDA device closure in a tertiary paediatric cardiac centre between March 2013 and March 2014 was performed. All children with PDA who underwent an echocardiogram within 24 hours of cardiac catherisation as part of transcatheter device closure of PDA were included in the study. Patients’demographic data were collected. Review of individual patients’ images from transthoracic echocardiograms using 8 MHz transducer linked with Phillips IE33 echocardiographic equipment recorded on MEDCON database was carried out. Measurement of PDA internal diameter on colour flow Doppler images from high left parasternal short axis view was performed at its pulmonary end in systole. Angiographic images recorded prior to transcatheter treatment of PDA in lateral view were reviewed on MEDCON database. Measurement of PDA internal diameter at its pulmonary end was performed. Krichenko’s classification of PDA type was used [7]. All PDA measurements were done independently by two observers on two separate occasions on mutually pre-agreed images. Both observers were blinded to the measurements of each other as well as their own previous measurements.

Descriptive data was presented as mean values ± standard deviation, or median values ± interquartile range if not normally distributed. We measured intra- and inter-observer correlation and agreement for both PDA internal size measurements on TTE as well as on angiography by calculating Pearson coefficient. We also compared PDA internal size measured on TTE with that measured on angiogram by calculating interclass correlation coefficient.

RESULTS

Fifty one patients underwent transcatheter device closure over the study period. Forty seven of these patients had adequate echocardiographic and angiographic data available for analysis. Median age was 7.5 years (range 8 months – 15 years). Mean weight of patients was 12.5 kg (SD 7.5). Demographic data and type of device used to close the PDA are shown in Table 1. None of the patients was found to have pulmonary hypertension significant enough to preclude safe closure of PDA using a coil or a device at the time of TTE or cardiac catheterisation. According to Krichenko’s angiographic classification [7] the following PDA morphology was encountered: type A – 66%, type E – 21%, type C – 11%. Majority of patients with history of prematurity had type C of PDA classification (71%).

Measurements of PDA size on TTE by a single observer on two separate occasions showed intra-observer correlation coefficient 0.5 (95% confidence interval 0.25–0.69). Measurements of PDA size on angiogram by a single observer on two separate occasions showed improved intra-observer correlation (correlation coefficient 0.75; 95% confidence interval 0.61–0.85). Measurements of PDA size on TTE by two observers on the same image showed inter-observer correlation coefficient 0.45 (95% confidence interval 0.18–0.65). Measurements of PDA size on angiogram by two observers on same image showed improved inter-observer correlation (correlation coefficient 0.83; 95% confidence interval 0.71–0.9). When the internal size of PDA on TTE was compared with measurement on angiogram poor correlation and agreement was obtained (Pearson’s coefficient 0.287 and interclass correlation coefficient 0.247) (Figure 1). The measurements of the same duct by the same observer from echocardiogram and angiogram were significantly different (p<0.0001).

DISCUSSION

Patent ductus arteriosus (PDA) is a relatively common cardiovascular condition seen in infancy particularly among population of prematurely born patients. Suspicion of PDA is usually raised by clinical examination, however, transthoracic echocardiogram (TTE) is the main diagnostic modality in clinical practice at present. PDA has been associated with a significant morbidity including chronic lung disease, intraventricular haemorrhage, retinopathy of prematurity, necrotizing enterocolitis in prematurely born neonates [4]. Various parameters seen on TTE including PDA diameter, ratio between left atrial and aortic valve size, blood flow pattern in superior vena cava, blood flow pattern across PDA, cardiac chamber size, and pulmonary artery blood flow pattern were described as useful tools in assessment of PDA haemodynamic significance in previously published studies [8, 9].

There are numerous studies published on the clinical significance of blood flow across PDA in premature neonates based on echocardiographic measurement of its size. Ratio between PDA size and size of the left pulmonary artery on TTE up to 4th day of life identified neonates less than 27 week of gestation who subsequently required closure of PDA [10]. Moreover, measurement of internal ductal size from colour flow Doppler images allowed for an early prediction of haemodynamically significant PDA in preterm infants and large ductal size showed a strong association with the decision to treat PDA (OR 4.3) [11, 12]. Effort to introduce a more sophisticated “PDA index” (PDA size²/weight in kg) showed that its value greater than 9 mm²/kg is unlikely to respond to medical treatment [13]. Furthermore, PDA size equal or above 1.5 mm or 2 mm measured on TTE in preterm neonates early postnatally was associated with increased odds of death, severe morbidity or considered an indication for surgical closure [14, 15]. Interestingly, a prospective study showed that PDA size measured on TTE in infants was uniform across a broad weight distribution indicating that PDA diameter measured on TTE did not vary significantly based on infants’ weight [16].

It is evident from the above studies that in many centres, especially in extremely premature infants, PDA diameter measured on TTE plays an important role in assessment and management of PDA. However, in significant number out of above studies, the exact method of measuring PDA size on TTE is missing making any direct comparison of their results difficult. There is also difference in methodology of measurement, where listed, as some centres use 2-dimensional images only while others 2-dimensional images enhanced by colour flow Doppler for assessment of PDA size. We visualised PDA in high left parasternal view, as distal continuation of the pulmonary artery space posteriorly into the descending aorta and measured its size at the pulmonary end on still colour flow Doppler images.

Attempts at measurement of PDA size using TTE have nearly 40 years long history [17] and given the advances in technology related to TTE, historical data is sometimes difficult to compare with recently published series. More recently (1998) Wong et al. tried to validate PDA diameter measured on colour flow Doppler TTE with angiography at the time of cardiac catheterisation [18]. This study included 28 children (median age 3.8 years, range 1 to 15 years) and showed that colour flow Doppler TTE significantly overestimated the true minimum PDA size. In animal study, Saunders al demonstrated that colour flow Doppler TTE often overestimated PDA size in dogs and transoesophageal echocardiographic data correlated better with ductal size measured on angiograms performed at cardiac catheterisation [19]. To our knowledge the above studies are the only ones published in English literature where attempts at validation of PDA size measured on TTE against angiograms were made. Intra- and inter-observer variations for PDA diameter measurements on TTE have not been published to date.

Our data show that there is a marked discrepancy between measurements of PDA size from TTE and angiograms. We failed to find clear evidence for overestimation or underestimation of PDA size on TTE when compared with angiographic measurements. Higher accuracy and reproducibility of measurements from angiograms is supported by better intra- and inter-observer correlation of measurements in our study. These results confirm the role of angiography as the gold standard in assessment of anatomical size of PDA. Given the retrospective nature of our study, we could not control the setting of colour flow Doppler range at the time of data acquisition. Differences in this setting could have influenced the accuracy of our PDA size measurements. We accept that our retrospective cohort of patients is relatively small and patients studied were older and bigger than the majority of patients in the so far published neonatal studies. However, it is somewhat disappointing that no unified method (e.g. standardised acoustic windows and views, probes’ resolution, place of measurement in the course of the duct, part of cardiac cycle when measurements should be made, unified setting of colour flow Doppler range) for the ductal measurements exists in the neonatal practice. Nevertheless, PDA size measured on TTE was used as eligibility criterion for randomisation in a placebo controlled trials of medical PDA treatment recently [20, 21]. Moreover, in the United Kingdom a multicentre randomised placebo controlled trial (OSCAR) of PDA treatment is currently under way where eligibility criterion for randomisation involves PDA diameter measured on TTE. The trial recommendation includes “gain optimised colour Doppler” used for PDA measurement, Doppler derived blood flow parameters, and measurement of PDA diameter at the site of maximum constriction.

CONCLUSIONS

Based on our data, angiographic measurement considered the gold standard for assessment of patent arterial ductal size does not correlate well with echocardiographic measurements. We therefore suggest that echocardiographic measurement of PDA size in preterm neonates should be interpreted with caution and any decision about treatment intervention should not be based on this measurement alone.

Došlo: 1. 12. 2016

Přijato: 19. 2. 2017

Corresponding author:

MUDr. Zdeněk Slavik, MD (UK), FRCPCH

Department of Paediatrics

Royal Brompton & Harefield NHS Foundation Trust

Sydney Street

London, SW3 6NP

United Kingdom

e-mail: Zdenek.Slavik@rbht.nhs.uk